Infrared and visible image fusion based on quaternion wavelets transform and feature-level Copula model

Xiaoqing Luo, Kai Li, Anqi Wang, Zhancheng Zhang, Xiaojun Wu

Abstract: To solve the problems that the correlation of multi-scale coefficients is ignored, as well as inaccurate identification of complementarity and redundancy of source images in traditional infrared and visible image fusion methods, an infrared and visible image fusion method based on quaternion wavelet transform (QWT) and feature-level copula model is proposed in this paper. The proposed method extracts the luminance, contrast and structure features of QWT magnitude and phase subbands. Then, the feature-level copula model capturing the inter-scale and phase-magnitude correlation is constructed to describe the intrinsic structure of images. The redundant and complementary feature type of QWT coefficient is further determined through the similarity of proposed models. According to the feature types, different fusion rules for high-frequency subbands are designed to accurately transfer the salient features of source images into fused image. For the low frequency subbands, a fusion rule is proposed using multiple features to avoid the degradation of image visual quality caused by false information. Finally, the fused low frequency subbands and high frequency subbands are transformed by inverse QWT to get the fused image. Experimental results show that the proposed method can effectively retain the rich details and structure information in infrared and visible images.